For following IBM encodings: x-IBM33722 and x-IBM964 that are listed in jdk1.5 documentation as supported encodings InputStreamReader misses every third character while reading them from FileInputStream char by char. I attached source of java application that demonstrates this problem and here's output from this application:


JIS_X0212-1990 failed. Pattern: '0123456789', line read char by char by InputStreamReader: '?????', line read by BufferedReader: '?????'

x-IBM33722 failed. Pattern: '0123456789', line read char by char by InputStreamReader: '0134679', line read by BufferedReader: '0123456789'

x-IBM834 failed. Pattern: '0123456789', line read char by char by InputStreamReader: '??????????', line read by BufferedReader: '??????????'

x-IBM964 failed. Pattern: '0123456789', line read char by char by InputStreamReader: '0134679', line read by BufferedReader: '0123456789'

x-JIS0208 failed. Pattern: '0123456789', line read char by char by InputStreamReader: '?????', line read by BufferedReader: '?????'

x-MacDingbat failed. Pattern: '0123456789', line read char by char by InputStreamReader: '??????????', line read by BufferedReader: '??????????'

Here are two supported encodings - x-IBM33722 and x-IBM964 for which BufferedReader and InputStreamReader read different lines:

"..line read char by char by InputStreamReader: '0134679', line read by BufferedReader: '0123456789'"

This problem reproducible on jdk1.5 (build 1.5.0_08-b03). In jdk1.6 (build 1.6.0-beta2-b86) it was fixed.